Intro to Viz

Stat 365: Statistical Communication

Monday, April 24th

Today we will…

  • Intro to Viz
  • History of Data Visualization
  • Intro to Tableau
  • Practice with Tableau

A taste of visualization

Dear Data1

U.S. Gun Killings in 2018

A big question for this course is how to best map variables to visual attributes1

A (brief-ish) History of Data Visualization

15,000 B.C.

Laxcaux, France

cave paintings1

900s1

1759 - 1823

William Playfair1

Credited with the invention of many common data visualizations: the pie chart, the bar chart, the line and area chart.

1813 - 1858

John Snow1

(no, not the one you’re thinking about)

Used mapping to solve a cholera epidemic in London

1781 - 1870

Charles Joseph Minard1

1820 - 1910

Florence Nightingale1

In addition to her work as a nurse, Nightingale was a statistician and invented the “coxcomb,” a variation on the pie chart

1822 - 1911

Francis Galton1

Super-famous statistician

and eugenicist

1800s

Statistical Atlases1

1868-1963

W.E.B. Du Boi1

(yes, the same Du Bois you’re thinking of)

Data Portraits of the Paris Exhibition

1897 - 1986

Mary Eleanor Spear1

Statistician who developed the box plot and bar chart!

1915 - 2000

John Tukey

  • Statistician who rocked the boat

  • Proposed a method called Exploratory Data Analysis (EDA), which involves making many simple graphs and summary statistics to understand data.

  • “The greatest value of a picture is when it forces us to notice what we never expected to see.”

  • Got credit for the boxplot, but didn’t create it

  • prim9: https://www.youtube.com/watch?v=B7XoW2qiFUA

1918 - 2010

Jacques Bertin

Cartographer and theorist

1943 -

William Cleveland1

  • Professor of statistics at Purdue
  • Did famous research about effectiveness of visualizations

Interactivity, brushing and linking

plotlyGGally from Carson Sievert on Vimeo.

Luke Tierney1

xlsp-stat

1945 -

Leland Wilkinson

Statistician and software designer

Worked on SYSTAT, SPSS, Tableau, now H2O.ai

1979 -

Hadley Wickham

Famous R programmer

Implemented the grammar of graphics in R, ggplot2

Works at RStudio

Mike Bostock

Created d3.js, a javascript implementation of the grammar of graphics

Things are accelerating!

Use a graph when…1

  • The message is contained the shape of the variables (e.g. patterns, trends, and exceptions)
  • The display will be used to reveal relationships among whole sets of values.

Anatomy of a Graph1

Let’s learn Tableau!

Data Connections

Tableau Sheet

Dimensions (think factors), Measures (think quantitative), and Marks (think aesthetics) are combined to create different charts to visualize data.

Continuous vs Discrete



Visualize the Guesses for Mandela’s Age by Anchoring Prompt

Download Mandela Data

Practice in Tableau

To do

Collect Data

  • due Thursday, 4/27

Data and Methods + Table

  • due Thursday, 5/4
  • You will need to include your statistical methods.
  • REPRODUCIBILITY! Document EVERYTHING!
  • Read CwD 3.4: Tracking the Analysis

One-number Story

  • Final submission: due Thursday, 4/27 at 11:59pm

Practice in Tableau

  • Submit URLs: due Monday, 5/1 at 11:59pm